Data Science Fellow - AI/NLP

We are seeking a postdoctoral researcher to develop AI/NLP and knowledge engineering methods that transform biomedical literature, experimental protocols, and source evidence into structured, quarriable, and evidence-grounded knowledge for organoid protocol standardization and optimization. 

The postdoc will work at the intersection of large language models, biomedical NLP, scientific document understanding, knowledge graphs, ontology grounding, computational biology, and human-in-the-loop curation. Potential projects include LLM-based protocol extraction, retrieval-augmented literature mining, curated knowledge graph construction, ontology and entity normalization, protocol comparison, consensus protocol derivation, benchmark design, and natural-language interfaces over structured biological knowledge.

Responsibilities

• Design and implement AI/NLP methods for biomedical literature mining and structured protocol knowledge extraction.

• Develop benchmark datasets, annotation guidelines, and evaluation pipelines for scientific information extraction.

• Build and evaluate RAG, in-context learning, fine-tuning, graph matching, entity normalization, and KG query workflows.

• Analyze extraction errors, model behavior, retrieval failures, grounding quality, and biological ambiguity.

• Collaborate with software engineers to integrate research methods into usable tools and reproducible pipelines.

• Collaborate with organoid biologists and domain experts to translate biological protocol knowledge into computable representations.

• Prepare manuscripts, conference abstracts, technical reports, design documents, and open-source research artifacts.

• Help define research milestones, evaluation criteria, and publication strategy for protocol intelligence work.

Required Qualifications

• PhD in computer science, computational biology, bioinformatics, biomedical informatics, NLP, machine learning, data science, or a related field.

• Strong Python programming skills.

• Demonstrated research experience with NLP, information extraction, LLMs, RAG, transformers, structured prediction, or scientific text mining.

• Ability to design controlled computational experiments, create benchmark datasets, and analyze results rigorously.

• Familiarity with biological, biomedical, or scientific data.

• Strong written communication skills and interest in publishing methods-oriented research.

• Comfort working with complex, evolving research codebases and interdisciplinary teams.

Preferred Qualifications

• Experience with scientific document processing, PDF parsing, biomedical literature mining, or methods-section extraction.

• Experience with knowledge graphs, ontologies, graph databases, graph algorithms, or semantic data modeling.

• Hands-on experience with fine-tuning LLMs, LoRA/QLoRA, Hugging Face, PyTorch, or API-based model evaluation.

• Hands-on experience with prompt engineering, structured JSON extraction, schema validation, tool use, or agentic LLM workflows.

• Hands-on experience with RAG systems, vector search, graph-augmented retrieval, or natural-language query over structured data.

• Exposure to bioinformatics concepts (e.g., sequence alignment, clustering, or phylogenetic analysis) that can inform protocol comparison and similarity methods.

• Background in stem cell biology, organoids, developmental biology, wet-lab protocols, or biological assays, enabling more effective collaboration with domain experts.